[
https://issues.apache.org/jira/browse/YUNIKORN-3092?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17986003#comment-17986003
]
John Daciuk commented on YUNIKORN-3092:
---------------------------------------
[~wwei] sorry, the "this line of code" link above was meant to point here:
[https://github.com/apache/yunikorn-core/blob/master/pkg/scheduler/objects/application.go#L853]
In other words, after DISABLE_RESERVATION="true", we still hit the log line
"reservation added successfully". This looks to be 1:1 with preemptions, so I
guess a pod that preempts still gets a reservation? I'm curious why that is.
Are we trying to guarantee that the preemptor can schedule to the node it
cleared?
Anyway, setting DISABLE_RESERVATION="true" immediately gives us a PASSING
preemption load test! We also put it through an ambient load test (just keep
throwing pods at yunikorn) and that looks good as well. So we deployed the
change to one of our user facing clusters today and we'll see how it goes this
week.
Regarding more details for our preemption test that fails without the env var,
I believe any two queues will do, as long as one should always be able to
preempt the other. To summarize the issue, if you have many pods all pending
that need preemption (say 100), in any scheduling round only 10 can attempt
preemption due to this line:
[https://github.com/apache/yunikorn-core/blob/master/pkg/scheduler/objects/queue.go#L48.]
However, it's likely that those 10 can't get a node because they are all
reserved by other pending pods. This leads to all the high priority pods stuck
pending even though there's tons of preemptible pods.
This preemption load test sounds unrealistic, but based on it it's not hard to
imagine how reservations can lead to problems in practice. Would be great to
get those problems sorted out so long as DISABLE_RESERVATION is not the default.
> Reservations can permanently block nodes, leading to preemption failure and a
> stuck scheduler state
> ---------------------------------------------------------------------------------------------------
>
> Key: YUNIKORN-3092
> URL: https://issues.apache.org/jira/browse/YUNIKORN-3092
> Project: Apache YuniKorn
> Issue Type: Bug
> Components: core - scheduler
> Affects Versions: 1.6.3
> Reporter: John Daciuk
> Priority: Minor
> Labels: preemption
> Fix For: 1.6.3
>
> Attachments: Screenshot 2025-06-23 at 8.28.55 PM.png
>
> Original Estimate: 336h
> Remaining Estimate: 336h
>
> h2. Context
> Since deploying Yunikorn back in October 2024 we've encountered occasional
> preemption misses. We find a high priority pod pending for hours, manually
> delete a low priority pod, then see the high priority pod schedule.
> We dug into this earlier this year and found the upgrade from 1.5.2 to 1.6.2
> helpful. In particular [this
> PR|https://github.com/apache/yunikorn-core/pull/1001] achieved 100% expected
> preemption in our testing due to it's reservation removal logic. However we
> still find that 1.6.2 is not reliable with respect to preemption in practice.
> And we can repro preemption misses by scaling up our original preemption load
> test by 4x.
> h2. Repro
> With Yunikorn 1.6.3 schedule ~400 low priority pods that live forever and
> fill up all node capacity. Once they are running, schedule the same number of
> high priority pods to a different queue. Use the same resources for all the
> pods.
> We expect that all the high priority pods will eventually schedule. However
> we find about 10% of them stuck pending. This can be seen in the screenshot
> attached, where the high priority pods are tier0.
> If we add logging like in [this diff from
> branch-1.6|https://github.com/apache/yunikorn-core/compare/branch-1.6...jdaciuk:yunikorn-core:jdaciuk-1.6]
> we see
> {quote}{{{}2025-06-23T04:54:26.776Z INFO core.scheduler.preemption
> objects/preemption.go:93 Removing node ip-100-76-60-239.ec2.internal from
> consideration. node.IsReserved: true, node reservations:
> map[847f05e1-f74c-403a-8033-154cd76d89c0:ip-100-76-60-239.ec2.internal ->
> tier0-1-395-157140|847f05e1-f74c-403a-8033-154cd76d89c0], node fits ask: true
> {"applicationID": "tier0-1-406-328120", "allocationKey":
> "e589c683-faf1-4793-97b8-c5f3b3bc34b5", "author": "MLP"{}}}}
> {quote}
> A node (with tons of potential victims) ip-100-76-60-239.ec2.internal is
> removed from consideration for preemption because it's reserved. Looking at
> the reservation map above, we see that pod tier0-1-395 has the reservation.
> The pod tier0-1-395 is blocking the entire node. Why can't it schedule and
> release the reservation?
> {quote}{{{}2025-06-23T04:43:45.942Z INFO core.scheduler.application
> objects/application.go:1008 tryAllocate did not find a candidate
> allocation in the node iterator, allowPreemption: true,
> preemptAttemptsRemaining: 0 {"applicationID": "tier0-1-395-157140",
> "author": "MLP"{}}}}
> {quote}
> Because there's no more preemption attempts allowed for the particular queue
> this cycle. And unfortunately this situation repeats itself every cycle since
> pod tier0-1-395 is not among the first in the queue to ever tryAllocate.
> h2. Thoughts
> This is one example, but there are a number of ways we can get stuck in such
> a cycle. Particular to the preemption failure here, it seems like we need
> some way to either remove the dead reservation or ignore it while considering
> preemption victims.
> So for example, when we iterate through the nodes in [this
> code|https://github.com/apache/yunikorn-core/blob/master/pkg/scheduler/objects/preemption.go#L163],
> perhaps we could first try with filtering out reserved nodes (as the code
> is) then try another loop ignoring and/or breaking reservations if we find
> victims. Would ignoring the reservation be enough, or do we have to delete it
> for the preemption to then result in scheduling?
> We'd love to get feedback as to the following
> * Is passing a test like described above even a goal of Yunikorn preemption?
> * If so, how can we be more strategic about releasing reservations that
> become major blockers, esp. in the preemption context?
> * We don't suppose there's a simple way to opt out of the reservation
> feature altogether is there? We don't ever want a reservation to block a
> node. If the pod can't schedule in the current cycle, we'd like it to wait
> without a reservation (in our case a full node will always free up at some
> point all at once). Or is there something we're misunderstanding that makes
> us need reservations?
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
